Gain Lab Department of Electrical and Computer Engineering University of Massachusetts, Amherst Course ECE 684: Microwave Metrology Lecture Gain and TRL labs In lab we will be constructing a downconverter. The above block diagram shows the basic components of this downconverter. The goal of the lab will be to characterize the gain of these components. Spring 2010 ECE 684 Microwave Metrology, Spring 2010 2 Image interference during downconversion Acos(! x t) A cos(! x t) cos(! LO t) = A/2 cos([! x -! LO ]t) + A/2 cos([! x +! LO ]t) Images in Downconversion image reject filter channel select filter cos(! LO t) IF component RF image IF filter selects channel! IF is the intermediate frequency. f LO f IF What RF frequencies could down convert to! IF? Either the upper side band! x =! LO +! IF or the lower side band! x =! LO -! IF would create the same IF component The undesired sideband is called the image frequency f LO Filter reduces image image rejection f IF RF and image downconverted to IF ECE 684 Microwave Metrology, Spring 2010 3 ECE 684 Microwave Metrology, Spring 2010 4
Calibration (S11) Calibration (S12) A good through standard will have a small return loss throughout the passpand Similarly, the insertion loss, postcalibration, should be zero ECE 684 Microwave Metrology, Spring 2010 5 ECE 684 Microwave Metrology, Spring 2010 6 Measuring Gain Image Reject Filter ZFL-1000LN A bandpass filter is used to reject images from the RF band. Note how markers are used to determine the insertion loss and image rejection. Amplifiers will provide signal gain. Most amplifiers are wide bandwidth. Care should be taken not to saturate the Network Analyzer when measuring S12 for amplifiers. This means that the stimulating signal should be less than -10 dbm. A special adjustment should be made on the network analyzer prior to calibration and measurement. ECE 684 Microwave Metrology, Spring 2010 7 ECE 684 Microwave Metrology, Spring 2010 8
Return Loss of a Bandpass Filter IF Amplifier Return loss for an ideal filter is 0 db (all power is returned), except in the passband. The IF amplifier is also broadband, but has a higher noise figure than the LNA at the front end. Note again the use of markers to quantify key measures of interest. ECE 684 Microwave Metrology, Spring 2010 9 ECE 684 Microwave Metrology, Spring 2010 10 Mixer The mixer is characterized by measuring the conversion loss (inverse of gain) and isolation of the ports, related to the desired signals. A Gain Table is Made for the Downconverter RF IF LO ECE 684 Microwave Metrology, Spring 2010 11 ECE 684 Microwave Metrology, Spring 2010 12
Gain versus Frequency Plot TRL Lab Purpose: Use the probe station to measure characteristics of discrete components that are mounted on a PCB. Probe L Compute the minimum and maximum frequency for the through line standars (20 deg min, 160 deg max) What is the image band? The pass band? ECE 684 Microwave Metrology, Spring 2010 13 ECE 684 Microwave Metrology, Spring 2010 14 TRL thru reflect (short) line Potential Problems DUT Impedances of lines must be known that s all Reference plane is center of thru Propagation constant comes out of calibration Error Sources Standards may be on ceramic substrate (alumina) different from wafer Pads on standards may be different than on DUT wafer Limited frequency range (extra line length between 20 and 160 degrees) Moving cables for line calibration On wafer standards take up space ECE 684 Microwave Metrology, Spring 2010 15 Calibrating with CPW standards removes probe related discontinuities up to reference plane set by TRL thru. DUT Calibrating with CPW standards and then measuring in microstrip (i) does not remove via discontinuity and (ii) improperly removes microstrip length. ECE 684 Microwave Metrology, Spring 2010 16
Potential Problems Potential Problems Common practice Better DUT reflect reflect Device to be measured is too close to probe. Coupling disturbes probe field and invalidates calibration Standard located at probe tips. Fields of probe and standard entangled Standard offset. Probe and standard only couple through CPW match match ECE 684 Microwave Metrology, Spring 2010 17 ECE 684 Microwave Metrology, Spring 2010 18 Potential Problems Through Standard Characteristics Probe tips too close. Reactive coupling changes during calibration Insertion Loss Structure allows a second mode to propagate Phase (S21) ECE 684 Microwave Metrology, Spring 2010 19 ECE 684 Microwave Metrology, Spring 2010 20
Through Standard Through Standard Return Loss Measured Attenuation Group Delay Simulation can be used to create a model for attenuation as a function of frequency (dependence on wavelength). Q: Why does attenuation go down at higher frequencies? Is this a real effect?? ECE 684 Microwave Metrology, Spring 2010 21 ECE 684 Microwave Metrology, Spring 2010 22 A 915 MHz Filter plotted using Ansoft Inductor Return Loss Q: Is there anything wrong with the measurement? What parts of the measurement behave as expected? Which do not?? ECE 684 Microwave Metrology, Spring 2010 23 ECE 684 Microwave Metrology, Spring 2010 24
Capacitor Return Loss A Model for the Inductor How about this measurement? How about this measurement? ECE 684 Microwave Metrology, Spring 2010 25 ECE 684 Microwave Metrology, Spring 2010 26 Values are found via simulations ECE 684 Microwave Metrology, Spring 2010 27